Competitive Intelligence for Deep Tech Companies
Find out which competing teams and technologies are really gaining ground, from their papers, patents, grants and pilots. Then choose your technical path, partners and pitch with evidence. The complete guide for founders, business developers and product leads at deep tech companies.
14-day free trial · 30-second setup · or read the guide
- 18 months
- how long a competitor's patent application usually stays secret European Patent Convention, Art. 93
Definition
What is competitive intelligence for deep tech companies?
Competitive intelligence for deep tech companies is following rival teams and rival technologies years before they sell anything. Their record is public: papers, patents, grants, pilots and hires. Reading it well tells you which technical path to back, which partners to sign first and what to tell investors.
The usual startup advice is to ignore competitors until one beats you with a shipped product. In deep tech, products ship years after the race starts, so the early signals come from labs, patent offices and grant portals instead.
Your strongest competitor may not be a startup like yours. It can be another physics: a different qubit, battery chemistry or chip design aimed at the same customer problem. It can also be the older technology, which keeps getting cheaper while you build.
Listed companies say so themselves. QuantumScape's annual report names the large cell makers CATL, LG Energy Solution, Panasonic and BYD as competitors, next to younger rivals like Factorial and Solid Power.
Use cases
Where competitive intelligence pays off in deep tech
Customers and investors compare you with every team chasing the same problem, often before you have a product. These are the jobs competitive intelligence does for you.
Reading technical claims
A record qubit count or battery density means little without the test behind it. Check the peer review, the benchmark and the test conditions before your team or your investors react.
Comparing approaches
The rival that wins your customer may use different physics. Compare what every approach must still prove, and how fast it has moved, rather than today's best numbers.
Mapping patent bets
Patent families show what a competitor protects and in which countries. Filings in China or Korea often point to manufacturing partners or customers there.
Following grants and programmes
Public funders publish a project's goal, partners, budget and end date. A funded project is a dated plan you can check against what the competitor later delivers.
Weighing pilots and partners
A pilot with a large customer is not a supply contract. Read the customer's own announcement next to the competitor's, and see how far the test really goes.
Watching who they hire
Process, test and quality engineers mean a competitor is moving from lab to production. Co-authored papers and conference talks show which labs it works with.
In practice
Competitive intelligence examples for deep tech companies
Deep tech competitors move through announcements long before products ship: a preprint, a grant, a pilot, a patent family. All of them deserve a checked answer, not a reaction.
Illustrative examples · CompetitorX is a fictional competitor
The competitor move
CompetitorX announces a record result in a press release, with a preprint.
Your move
Read the preprint, not the headline: the test conditions, the benchmark and what was left out. Ask your scientists for one plain paragraph your team can give to customers and investors who ask.
Before your next investor or customer call.
The competitor move
CompetitorX's EU-funded project names your target customer as a partner.
Your move
The project page lists the partners, the budget and the end date. Ask the customer what the project covers. Then offer a test on a problem the project leaves out.
Within the month.
The competitor move
CompetitorX files patent families in China and Korea in your area.
Your move
It may be lining up manufacturers or customers there. Ask your patent counsel whether the filings touch your plans, and check which countries your own filings cover.
Within the quarter.
The competitor move
CompetitorX announces a pilot with a car maker you are talking to.
Your move
Read the car maker's release as well. A test with one unit is not a supply deal. Keep your own conversation going on what the pilot does not test, such as cost at volume.
That week.
The competitor move
A Chinese manufacturer offers a price 40% below your target cost.
Your move
Find out what the price covers: volume, quality, qualification, delivery. Then show customers where your cost falls as you scale, and what they get that the cheaper product lacks.
Before your next customer review.
The competitor move
A government programme selects CompetitorX for its next stage, and not you.
Your move
Read what the programme says the stage tests. Explain to your board what the selection proves and what it does not. Then show your own evidence on the same questions.
Before your next board meeting.
What to know
What to know about each deep tech competitor
Patent offices, grant portals and preprint servers answer most of these questions. Customers, partners and your own scientists fill in what those records leave out.
Their approach
- Which technical approach do they use, and what must it still prove?
- Which milestones have they promised, and by when?
- Which promised dates have they already missed?
- What would a customer have to change to use their product?
Their evidence
- Which results are peer-reviewed, and which are press releases?
- Which benchmarks do they report, and who defined them?
- Has a customer or an independent lab tested the product?
Their backers
- Which grants and programmes fund them, and until when?
- Which customers run pilots, and how do those customers describe them?
- Who will manufacture for them, and in which country?
- How long does their last funding round have to last?
Their people
- Which specialists are they hiring now?
- Which labs and universities do they publish with?
- Which of their scientists speak at which conferences?
Sources
Where deep tech companies find competitive intelligence
Deep tech competitors publish their work long before they sell it: in journals, patent offices and grant portals. That record is free, but slow and written for specialists.
What you already hear
- Customer and partner meetings
- Industrial customers often test several approaches at once. Ask which others they are evaluating, and on which criteria they will decide.
- Your scientists
- Your researchers read competitors' papers anyway. Agree on how they flag a result that should change your plan or your pitch.
- Investor meetings
- Investors' questions often echo a competitor's pitch from the week before. Write those questions down after every meeting.
- Conferences
- Your team hears competitors' talks and sees their posters. Log what was shown working, not only what was promised on stage.
What competitors publish
- Preprints and papers
- arXiv offers an RSS feed per field, and Google Scholar sends alerts on an author's name. Check the journal's page too: editorial notes and corrections live there.
- Patent databases
- Espacenet, WIPO Patentscope and Google Patents are free. Most applications stay secret for 18 months, so read competitor patents as a record of past work and of the countries a rival targets.
- Grant databases
- CORDIS lists EU projects with partners, budgets and goals. UKRI's Gateway to Research and SBIR.gov do the same for UK and US awards. Grants are the part of competitor funding that comes with a written plan.
- Programme selections
- Government programmes publish who passes each stage. In November 2025, DARPA's quantum programme named 11 companies for its second stage, six of them from Canada, Australia or the UK.
- Annual reports of listed rivals
- Listed companies like IonQ, Rigetti and QuantumScape must describe their competitors and risks. Their SEC filings also date the milestones they promised investors.
- Customer announcements
- Read the customer's press releases, not only the vendor's. The customer's words show whether it is a lab test, a pilot or a purchase.
- Job ads
- Roles in process, test or quality engineering show a move from lab to factory. Job postings also name the tools and materials a team works with.
- Export control lists
- EU and US export rules now cover quantum computers, cryogenic parts and chip equipment. They decide which competitors can sell to which countries.
Stay on the right side of the line
Papers, patents and grant pages are published to be read, so read them freely. Never pose as a customer, partner or investor to see a competitor's lab, and never quiz a new hire about a former employer's results, processes or unpublished data. Leave any judgement on whether a competitor infringes your patents to counsel, and keep it out of competitive documents. Check export rules before you share technical data with a partner abroad.
Signal vs noise
Deep tech competitor news: what counts and what doesn't
Deep tech news runs on announcements: records, roadmaps, rounds and demos. Act on what has been checked and could change a customer's or an investor's decision.
Track
Act within a week
- A peer-reviewed result in your area
- A grant or programme stage won
- Patent families in your target countries
- A pilot the customer confirms
- Hiring for production roles
Skim
Monthly roll-up
- Preprints
- Funding rounds
- Roadmap announcements
- Conference talks
- Partnership press releases
Ignore
Unless it repeats
- Records with no test conditions
- Demo videos
- Valuation headlines
- Market-size reports
- Arguments on social media
Preprints are the hardest call: often real, but not yet checked. Microsoft's 2025 Majorana paper in Nature carried an editorial note saying its results "do not represent evidence for the presence of Majorana zero modes". The press coverage had moved on long before.
Funding rounds sit in the skim column on purpose. A large round buys a competitor time, not results. Note how many years it must cover, then see what it buys: hires, a pilot line, a factory.
See what deep tech competitors change online
Flares watches competitors' websites, product pages, news and hiring, and tells you every week what changed.
14-day free trial · 30-second setup
Distribution
How competitor knowledge moves in a deep tech company
In a deep tech company, scientists read the competitors' papers while the business team hears what customers test. The two rarely compare notes until a board meeting forces it.
What comes in
Research and engineering
Competitors' papers, patents and results, read by experts.
Business development
Which approaches customers are testing, and on which criteria.
Grants and funding
Who won which programme, with which partners.
Recruiting
Who competitors hire, and who leaves them.
What goes out
LeadershipTechnical strategy review
Which approaches gain ground, and what it means for yours.
Board and investorsBoard update
Competitors' rounds, milestones and missed dates, checked.
Business developmentBriefs
Plain answers to the competitor claims customers repeat.
ProductSpecification review
Targets set against competitors' verified results.
Patent counselPatent review
Competitor filings that touch your plans.
The weakest link usually sits between the lab and the business team. A scientist knows a competitor's record is shaky, and the person in the customer meeting doesn't. Ask the lab for one plain paragraph on each major claim, written for a non-specialist.
Founders usually carry this until the first commercial hire. Product managers then turn it into specifications, and your venture capital backers will ask about the same claims at board meetings.
The deliverable
What goes on a deep tech competitor evidence sheet
Investors and customers will ask how you compare with every rival team. Give them one page per competitor that separates what it claims from what it has shown, with a source for every line.
01Approach
The technology, and what it must still prove to reach customers.
02Claimed results
Each claim, its source, and whether it was peer-reviewed.
03Benchmarks
The metrics reported, who defined them and the test conditions.
04Promised dates
Each milestone with its date, and the ones already missed.
05Patents
Families by technical area and country, with priority dates.
06Grants and programmes
Projects, partners, budgets and end dates.
07Pilots and customers
Each pilot, in the customer's own words.
08People and money
Key hires, the last round and how long it must last.
Place each sheet on a competitive landscape map of approaches, so the team sees where rivals stand and which path is crowded.
When you need numbers side by side, a competitive benchmarking table keeps every value's test conditions next to it.
The deep tech calendar
The deep tech year: conferences, grant rounds and decisions
Deep tech moves on the science calendar. Competitors time their results and announcements around the big conferences, grant deadlines and programme decisions.
- 1
Chip conferences
February and August- ISSCC in February and Hot Chips in August show new chip designs.
- Compare the numbers shown with the claims made before.
- 2
Spring science and industry
March and April- The APS March Meeting brings quantum and materials results.
- Hannover Messe shows robots and industrial hardware.
- 3
EU grant cut-offs
Six times a year- The EIC Accelerator had six cut-offs in 2026, from January to November.
- Results arrive months later, so check who won in your area.
- 4
Quantum and climate events
September to November- IEEE Quantum Week brings quantum results in early autumn.
- The UN climate conference in November draws climate hardware news.
- 5
Programme decisions
Set by each programme- Government programmes name who passes each stage.
- Read what a stage tests before anyone quotes it.
- 6
Patent publications
Every week- Applications appear about 18 months after the first filing.
- Alerts on competitors' names catch them as they land.
Routine
A deep tech competitor routine
Deep tech changes slowly, so the routine stays light: a short weekly scan of your alerts, and a deeper read when a big result or decision lands.
Weekly
15 minutes- Scan preprint and patent alerts for competitors' names.
- Note any new claim, with its source.
- Flag what customers or investors will ask about.
Monthly
1 hour- Check grant databases and programme news.
- Read competitors' new ads for specialist roles.
- Update the list of promised dates.
Quarterly
Half a day- Review patent families by area and country.
- Update the evidence sheets and the map of approaches.
- Tell the board what changed, and what did not.
After a big claim
A day- Read the paper or preprint in full with your scientists.
- Write a neutral paragraph for customers and investors.
- Decide whether it changes your plan.
If you summarise a long paper with an AI assistant, a source grounding prompt keeps every statement tied to a page of the paper, so nothing gets invented.
Freshness
How long deep tech competitor intel holds
A record result can be beaten within months, while a technical approach lasts for years. Use this table to time your rechecks.
| What you track | Goes stale in | Update it when |
|---|---|---|
| Claimed records | Months | A peer review, a correction or a better result |
| Technical approach | Years | A change of approach or a new product line |
| Promised milestones | A quarter | The promised date passes |
| Patent families | A quarter | New publications or a status change |
| Grants and programme stages | Six months | A new award or stage decision |
| Pilots and customers | A quarter | A customer announcement |
| Manufacturing partners | A year | A new supply or factory deal |
| Cash and funding | Six months | A round, a registry filing or layoffs |
| Key scientists | A quarter | A move, a paper or new job ads |
| Export rules | A year | A new control list |
| What they tell investors about you | Weeks | An investor repeats a new claim |
Keep the missed dates on the sheet. A competitor's record on its own deadlines is the best guide to its next one.
Metrics
How deep tech teams measure competitive intelligence
With a handful of deals a year, count rather than calculate rates. Then add how fast you answer a competitor's claim and how current your evidence is.
Competitive win rate
won competitive deals ÷ (won + lost competitive deals)
Count pilots and design partner choices as deals. With so few, report the counts: "three of five against CompetitorX" says more than 60%.
Time to insight
median days between a competitor's announcement and a checked answer for your team
Measure it after every big claim. Investors read the same headline you do, often on the same morning.
Intelligence freshness
items verified in the window ÷ items requiring verification
Apply it to the evidence sheets. A sheet with last year's results is the one a customer will correct for you.
Keep a decision log too: each time competitor evidence changed your technical path, a partner choice or your pitch. In deep tech, that log shows the value of the work better than any rate.
Pitfalls
Competitive mistakes deep tech companies make
Deep tech rewards patience and punishes hype, in both directions: believing a competitor's headline, or overselling your own.
Comparing headline numbers
Qubits, watt-hours and tokens per second are measured in different ways. Compare like with like, or not at all.
Believing the press release
Wait for the paper, the customer's own words or an independent test before you change anything.
Ignoring other approaches
The rival that beats you may use different physics. Track every approach that solves your customer's problem.
Forgetting the old technology's cost
Older technology keeps getting cheaper. Your target cost has to beat its future price, not today's.
Treating a grant as a win
A grant buys a team time to try. Watch what the funded project actually delivers, and by which date.
Overlooking Chinese manufacturers
They often reach volume faster and at a lower cost. Follow their prices and output, not only Western startups.
Automation
How to automate competitive intelligence for deep tech companies
Patents, papers and grants come with their own alerts. What slips is the competitor's own website: a new product page, a changed specification, a partner's logo, a careers page full of process engineers.
Competitive intelligence platforms are made to watch those pages. Flares follows competing deep tech companies' websites, product and specification pages, news, messaging and hiring, and sends each change with what it may mean for you. If you record customer calls in Gong or Modjo, the competitors customers mention there are included too. It does not read patent offices, grant portals or preprint servers. Keep those alerts in your weekly scan.
Automated competitor monitoring
Competitors' product pages, specifications, news and careers pages, with changes flagged.
Weekly competitive digest
Last week's changes at competing deep tech companies, ready for your Monday meeting.
AI competitive analysis reports
One competitor's products, partners and messaging in a single report, ready for a board pack.
Answer investors' deep tech competitor questions fast
Flares keeps a current view of each competitor's products, partners and hiring, ready for your next board pack.
14-day free trial · 30-second setup
FAQ
Deep tech competitive intelligence FAQ
What is deep tech?
Deep tech companies build products on new science or engineering: quantum computers, robots, new chips, materials, batteries or fusion. They spend years and large sums before revenue, and their main risk is technical. WIPO's 2026 Global Innovation Index counts more than 30,000 science-based startups created since 2000.
Is artificial intelligence deep tech?
AI chips and robots are deep tech, and most AI software is not. A new processor or a robot takes years of hardware work. An app built on someone else's model can be copied in weeks, so AI software companies compete more like SaaS companies, on price, features and distribution.
What is competitive intelligence for a deep tech company?
It is the ongoing work of following rival teams and technologies before they sell: their papers, patents, grants, pilots and hires. Founders and business development leads use it to choose a technical path, pick partners and answer investors' questions about the competition.
Can you give an example of competitive intelligence in deep tech?
When DARPA named 11 companies for the second stage of its quantum programme in November 2025, every quantum startup had a new public fact to explain. The useful reading came from DARPA's own words: the programme says it is "not a competition between performers" but a test of each approach.
How do you do a competitor analysis for a deep tech company?
List every approach that could solve your customer's problem, not only startups like yours. For each rival, record its claimed results with sources, promised dates, patents, grants, pilots and key hires. Then rank them on what each still has to prove. A competitor profile template gives all rivals the same layout.
What are the four main types of competitors in deep tech?
Direct competitors use your approach for the same customers. Indirect ones use different physics for the same problem. The replacement is the older technology the customer already runs. Potential competitors are large companies' research labs and manufacturers that could scale a rival approach quickly.
Which framework works for a deep tech competitor analysis?
Technology readiness levels give everyone the same scale. The EU defines level 4 as "technology validated in a lab" and level 9 as "actual system proven in an operational environment". Place each rival on that scale, then add Porter's five forces for the market around it.
How can you tell if a competitor's technical claim is real?
Check four things: peer review, the benchmark, the test conditions and an outside test. "They're quoting 400 watt-hours per kilogram, but nothing here is telling me what the masses or volumes are," said Eric Wachsman, a battery researcher, of a 2026 claim. Missing conditions are the most common gap.
What do a competitor's patents tell a deep tech company?
They show what the competitor worked on about 18 months ago, and where it expects to sell or manufacture. A family filed in China and Korea points at Asian partners. They don't show what will ship, since many patented ideas never become products.
Can you find out which grants a competitor won?
Usually, yes. CORDIS lists EU-funded projects with partners and budgets, UKRI's Gateway to Research covers the UK, and SBIR.gov and NSF list US awards. Competition for them is tough: at the European Innovation Council's October 2025 cut-off, 61 of 923 applicants were funded.
How do you compete with a better-funded deep tech startup?
Pick the customers and problems its approach serves worst, and prove your result there first. Money decides who reaches the next stage, though. "There's a Darwinian selection at each phase," said Francesco Sciortino of Proxima Fusion in 2026. Track what the rival's money buys, not the size of the round.
How do you compete with Chinese manufacturers on cost?
You will rarely win on price alone. In 2025, battery packs in China cost 30% less than in North America and 35% less than in Europe, according to the IEA. Compete on what scale cannot buy quickly: performance, qualification with local customers, supply security and export rules.
Is it legal to hire a competitor's scientists?
In most countries, yes, if they bring their skills and not their former employer's documents. Don't ask for lab notebooks, process recipes or unpublished results, and check their non-compete and confidentiality terms before they start.
Is competitive intelligence legal for deep tech companies?
It is, as long as you read what is published and never pretend to be someone else. Papers, patents and grant pages are public. Don't pose as a customer or investor to see a competitor's lab, and leave infringement questions to counsel.
Can ChatGPT do market research for a deep tech company?
It can summarise papers and patents quickly, but it misses unpublished work and can invent numbers. Give it the documents themselves, ask for a page reference for each fact, and check each figure against the source before it reaches a board pack.
What tools do deep tech companies use for competitive intelligence?
Free alerts cover most of the record: arXiv feeds, Google Scholar, Espacenet or Patentscope, and the grant portals. Competitive intelligence software adds what those miss, the changes on competitors' own websites, news pages and careers pages.
Ready? Your competitors won't wait for you.
Get your first competitive digest next Monday.
14-day free trial · 30-second setup