Raising the Bar: Inception-Stage Venture in the Age of Consensus

TL;DR: Venture capital has bifurcated into two extremes, and inception-stage funds have to be willing to play both—while recognizing that 99% of the expensive consensus market is fool's gold at pure gold prices. For the true "N of 1" outlier, paying up makes sense. But the vast majority of our capital, bandwidth, and conviction belongs on the non-consensus side. Navigating that reality with Series A graduation rates way down means rewriting our priors and rewriting seed portfolio math: more initial shots on goal (30-40 instead of 20-24), ruthlessly cutting defensive follow-on reserves, and underwriting businesses that can survive outside the consensus spotlight.

This is a confusing and dangerous time to be a venture investor. It’s also, however, one of the most exciting times to approach the finance of inception-stage innovation. We are re-inventing entire industries and tech stacks overnight, companies are growing faster than ever before, and (the siren songs of the AI doomers aside) we are living through an incredibly optimistic time - one in which anything and everything feels possible all at once.

There are three critical concepts that appear to be defining this era in venture investing: acceleration, diffusion, and concentration.

Acceleration. Technology feels like it is moving faster because, in fact, it is. AI has enabled founders and engineers to ideate faster, build faster, learn faster, and grow faster than ever. We see this directly in AI itself but also in downstream technologies such as software engineering, mechanical engineering, and biotech - where AI is enabling an ever increasing pace of invention. That’s a great backdrop for investing in innovation. It’s never been easier or more exciting to build.

Diffusion. Innovation is happening worldwide and in every possible vertical or niche industry. As an inception-stage fund operating in Europe and Israel, we have a front row seat to this diffusion, and we can confirm that the process of diffusion is only intensifying. An increasingly diverse set of founders is going after an increasingly diverse set of very interesting challenges. The market size of the innovation eco-system keeps getting bigger.

Concentration. Any observer of the venture landscape over the past two years has observed the same set of trends: capital is concentrating at unprecedented levels. While the total amount of capital deployed in the venture world has skyrocketed, this is getting concentrated into an ever smaller number of venture firms, tech companies, and ideas. Entry rounds and exit outcomes that defy gravity have acted as a lightning rod for investor imaginations, enabling the right founders in the right spaces at the right moments to raise increasing amounts of capital at previously unimaginable rates.

The implications of acceleration and diffusion are clearly positive for founders. In absolute terms, it’s just a great time to build. With or without venture capital backing, the opportunity for founders and builders to make a major impact on a problem they care deeply about has never been larger. The problems for founders and the VCs who love them begin with concentration.

Same but different. Firms like Angular exist to fund companies at the start of their journey. In a lot of ways, our work in 2026 remains identical to what it was in 2019 when we got started or, frankly, to what it would have looked like in 1985. We’re looking to back outstanding people doing unlikely things that could be huge if they work and have some plausible pathway to pseudo-monopoly status with high growth and high margins. That’s inception-stage venture in a nutshell, and, no, I don’t think the essential nature of it has changed dramatically despite the fact that we can get to meetings in self-driving cars and there are AI notetakers on some of our calls. What’s dramatically changed, however, is the downstream financing market - and we can’t ignore that reality. A founder or a company that is not “legible to capital” (see this piece by Nikunj Kothari) is going to struggle to raise capital in the current climate.

Last week, Ethan Kurzweil of Chemistry Ventures wrote a powerful piece that summed up his firm’s posture. I agree with it wholeheartedly, and want to quote it at length before diving into the portfolio construction implications of what he’s written:

"It’s important to bear in mind that these two markets are not mutually exclusive…we’ve backed (pricey) consensus AI founders as well as those building in sectors as far from the heat as the dark side of the moon. And if past is prologue, there will be gems—and false gold—in both camps. But we are clearly still early to this hyper-consensus market, so timing is not irrelevant to the equation. As a result, our collective advice for founders has been to shift to the extreme poles – if you’re out of favor, prepare for life in that lane for longer than you think and make sure your business (and psyche) can survive there. And if you’re more of a consensus needle mover, that’s not a train you can get off and rest up before the final destination. In the end, future Trilicorns will have lots of different origin stories – but being blind to the reality and implications of the market is no longer a viable path."

Ethan’s nailed it. There are two viable paths for an inception-stage company and, therefore, two distinct paradigms of inception stage investing that make sense for a firm like Angular (or Chemistry) to pursue. The first (non-consensus) has become dramatically harder and has three sub-categories within it. The second (consensus) has emerged in recent years, threatens to destroy every venture firm that touches it, and is far rarer than anyone realizes.The market has barbelled into consensus haves and non-consensus have notes, and our portfolios are likely to do the same to some extent.

Inception-stage consensus investing is a viable but vanishingly narrow path. Let me start with consensus investing. The key here is to distinguish between pure gold and fool's gold. We’ve all seen the decacorn exits and the trillion-dollar IPOs of the AI era. We all know these outcomes are possible. Over the past several years, entry valuations for every round type in every sector in every geography have crept up - and there is always a bigger and more expensive seed round lurking in next month's headlines. While it’s generally good advice not to pass up on a “generational opportunity” because of price, the reality is that every remotely consensus opportunity is being priced to perfection. There are “N of 1” founders out there, and when a seed fund (including Angular) sees one, it should invest. We’ll pay that very high price in those tiny handful of cases where doing so makes sense because we are convinced the founding team is truly pure gold “N of 1” and (crucially!) that the downstream fundraising challenge (the central obstacle to exit viability in this era of consensus) has already been overcome. The crucial thing, however, is to remember that the majority of companies that raise expensive inception rounds are fool’s gold, not pure gold and not gold at all. Just a shiny object glinting in the sun and drawing in a ton more capital than it deserves. The vast majority of the venture market is systematically paying pure gold entry prices for fool’s gold companies. The bar for “paying up for quality” has never been higher.

The world as it is: non-consensus venture is more valuable than ever but we’re all gonna chew glass. Non-consensus inception investing is the core of what we do at Angular Ventures. The mispricing here is astounding, and the opportunity to create value through this type of investing has never been greater (acceleration and diffusion), but it’s also never been harder to secure follow-on financing. There is extensive data showing that “Seed to Series A” graduation rates have been plummeting since 2021 and that the bar for a Series A continues to rise across every metric (ARR, growth rate, NRR, you name it). As a result (and as Nnamdi Iregbulem from Lightspeed has painstakingly documented) the population of seed-funded companies is dying off. Being an inception-stage founder in a non-hot category has never been harder even as innovation and execution have never been easier. For VCs, the implications are several:

1. First, we need to raise the bar. Venture is so hard that this is always a good idea - but especially now. With graduation rates as low as they are, venture investors at the seed stage need to make sure they never lose sight of the sheer difficulty and low probabilities of what they are doing. For hands-off investors, this might be easier. For investors (like us) who aspire to be useful board members and thought partners, this is demanding. For founders, it's brutal.

2. Second, we need to be extremely aware of what risk we are underwriting and what outcome we are anticipating. There are three flavors of non-consensus investing: fund to breakeven (seedstrap), fund to fundraise (become consensus), and fund to fund again (internal milestone). All are viable. All are incredibly risky. In this climate, especially, however, miscategorizing an opportunity is a recipe for disaster. Underwriting a “fund to fundraise” bet as if it were “fund to breakeven” leaves the company undercapitalized, while assuming a non-consensus bet will magically graduate to consensus leads to an early grave. The good news is that in the world we’ve entered - all of these pathways can lead to asymmetric power-law returns for appropriately sized inception-stage funds.

3. Third, we need to communicate better and more clearly with founders. We’ve entered into a no-nonsense era of inception-stage investing. The consequences of missing a plan or making the wrong bet have never been more dire. In other times, the Series A graduation rate meant that many of the mistakes of the seed stage were papered over, and the founders got a mulligan with more capital. Today’s world is far less forgiving - and requires open and honest communication between founders and funders.

4. Fourth, we must be resilient and patient. Hard things take time and often don’t move linearly. Today’s unloved non-consensus company can become tomorrow’s VC darling or next week’s profitable high-growth miracle. But it takes time and resilience.

Some thoughts on portfolio construction. The portfolio implications of this are increasingly clear: it’s the basics of inception-stage just ramped up a notch higher. First, and most importantly, we need more shots on goal given the dramatically higher systemic risk. A seed fund that could have done 24 investments and expected a 50% graduation rate should probably do 40 investments and expect a 10-25% graduation rate. Second, the bar for follow-ons must be even higher. The justification, price, and signal around every follow-on check needs to be looked at in incredible detail - and every follow-on needs to be weighed against the opportunity to have placed another shot on goal at inception. Companies that raise at steadily higher valuations may not carry with them the same signal value as they used to. Companies that manage to successfully raise at sky-high “consensus” valuations may no longer present an attractive risk-reward profile for new dollars despite being a rising star. In practice, this means actively shifting fund reserves away from reflexive or defensive follow-ons and reallocating that capital toward more initial checks and sniper-shot follow-ons that make economic sense.

Raising the bar. Across this analysis - and across my thinking in recent months - raising the bar is the one constant. In the bulk of our work - leading non-consensus inception rounds - we need to raise the bar. In those handful of cases where we believe we have access to a pure gold “N of 1” founder, we need to raise the bar. When it comes to reserves and follow-ons we need to raise the bar. As Jerry Colonna writes, what if the real work is learning to hold contradictory truths at the same time? This is indeed one of the hardest times to be a venture capitalist or a founder. At the same time, it is one of the best. 

Gil Dibner

Could the future of software be fluid
How do we get the best of AI without losing the soul of software?

The future belongs to young missionary teams
Why it makes more sense betting on youth in the current moment

The AI-native enterprise playbook
Ten real-time observations on a rapidly evolving playing field

No more painting by numbers
It’s the end of the “SaaS playbook.

HARD TECH

A new scaling law? OpenAI claims to have solved the Navier–Stokes Millennium Prize Problem using a model “significantly more capable” than GPT‑6 Astra, roughly 10,000 concurrent agents, 130 billion output tokens and a further 17 hours of formal verification in Lean. If the proof survives expert scrutiny, the result suggests that the next frontier may come not only from better models, but from spending enormous amounts of inference compute to coordinate thousands of agents against a single valuable problem.

The frontier asks for a speed limit. Anthropic CEO Dario Amodei argues that recursive self-improvement and recent examples of misaligned agent swarms mean AI capabilities are now advancing faster than safety measures can keep up, calling for embedded third-party evaluators, industry coordination and eventually global agreements. This means a leading frontier lab believes the race has become too dangerous to govern through corporate self-restraint alone, even as competitive and geopolitical pressures make slowing down individually nearly impossible. Sam Altman and Elon Musk both noted their agreement on X, suggesting that the top three labs (except Google) may want to slow down together.

Humanoids get an annual upgrade cycle. Unitree has launched the $15,000 G1+, adding a two-axis moving neck, binocular vision and stronger motors that lift maximum arm load from roughly two to three kilograms—all for an 11% premium over the G1. The rapid, relatively inexpensive improvement of Chinese humanoid hardware strengthens the case that bodies will commoditize faster than expected, shifting the most valuable bottleneck toward the software, data and autonomy required to make them reliably useful.

HARD MARKETS

The customer is the investor. The Boring Company has raised $3 billion at a $23 billion valuation, led by the UAE, which has also committed to more than 150 kilometers of tunnels despite the company having opened a passenger system only in Las Vegas. The arrangement is a useful model for capital-intensive startups: pair financing with a sovereign customer able to supply capital, demand and permitting—but the fourfold valuation increase also shows how readily a giant future contract can be priced as though execution risk has disappeared.

Loyal wingmen take flight. Anduril’s autonomous YFQ-44A Fury has now flown alongside F-35s, another step toward the Air Force’s plan to pair crewed fighters with cheaper uncrewed aircraft carrying additional sensors and weapons. The milestone is bigger than a successful flight test: a startup-built combat aircraft is being integrated directly with one of America’s most advanced weapons platforms, showing how quickly new defense companies can move from supplying components to competing for the military’s core programs.

HOW TO STARTUP

The emperor’s dilemma. Jeremy Stern’s extraordinary profile of Mark Zuckerberg portrays a founder still driven by the simple compulsion to build, but increasingly aware that unmatched power and decades of success can isolate him from reality. It also recasts Meta’s many expensive detours—from phones and crypto to the metaverse and now AI glasses—as repeated attempts to escape Apple’s platform control, capturing both the strategic advantage and profound danger of giving a singular founder enough authority to pursue a 20-year vision.

Distribution strikes back. Meta’s new consumer agent Muse has climbed to No. 2 in the U.S. App Store with more than 83,000 iOS downloads, although its launch remains far smaller than Threads, Meta AI or ChatGPT. The early numbers suggest that Meta’s enormous distribution and WhatsApp integration make it an immediate contender, but not an automatic winner—leaving room for startups like Instinct to build a new social graph around agents that coordinate with one another on their users’ behalf.

HOW TO VENTURE

Pacing gets momentum. Zvi Mowshowitz argues that Dario Amodei’s proposal has already produced “actual progress”: OpenAI matched Anthropic’s commitment to embedded external evaluators, while Elon Musk and Demis Hassabis endorsed the broader call to slow frontier development. The unresolved question is how anyone verifies that progress has truly slowed rather than merely becoming more expensive—which means the evaluators, standards and measurement systems now being designed could become the first real regulatory architecture for the frontier-model industry.

The venture barbell gets sharper. Nicole DeTommaso argues that mega-funds are squeezing traditional lead investors out of competitive rounds, while longer liquidity timelines and high consensus pricing make mid-sized fund economics increasingly difficult. Her prescription is to run a smaller fund, invest at true pre-seed before consensus forms and preserve the flexibility to sell some secondary shares—another sign that the durable venture models may be giant multi-stage platforms at one end and focused inception investors at the other, with the middle increasingly hard to defend.

Steadybit's MCP Server has been reviewed by Anthropic and is now live in the official Claude directory.

CruxOCM
Advanced Process Control Engineer (Calgary, Vancouver, Remote)

FalkorDB
Cloud Architect (Tel Aviv, Hybrid)

Sourcix
Full Stack Developer (Tel Aviv)