Conviction and Similar Pre-Seed AI Investors: Complete Comparison Guide (2026)

Founders comparing AI-focused pre-seed investors usually start with one name: Conviction. The follow-up question is the same most of the time: who else looks like this? Conviction sits in a small but real archetype, partner-led and AI-native, writing $250K to $5M into seed and Series A, with a separate program for pre-seed. This guide breaks down what Conviction actually does, what “similar” looks like in 2026, and which kinds of investors operate alongside it at pre-seed.

Should you pitch Conviction directly or apply to Embed first?

Three questions sort this out before any other reading.

Are you at seed or later? If you have $1M+ in ARR or strong AI infrastructure traction, you can pitch Conviction’s main fund directly. That’s where their core check sizes ($1M to $25M) and stage focus sit.

Are you at pre-seed? Direct pitches to Conviction’s main fund at pre-seed are unusual. The intended entry path is Conviction Embed, the firm’s dedicated cohort program. Embed runs roughly twice a year. The 6th cohort closed for applications on February 23, 2026, per the Embed website (embed.conviction.com).

Are you outside Conviction’s actual scope? If you’re not building AI-native software, or your company is not at seed/Series A and you missed the latest Embed cohort, look at the structural archetypes in section 4. Several types of comparable investors operate on different timelines.

What does Conviction actually invest in (and not invest in)?

This is the entity disambiguation most lists skip.

Conviction is the venture firm founded by Sarah Guo in October 2022, after she left Greylock Partners. The firm is based in San Francisco. Mike Vernal joined as second General Partner in late 2024 alongside Fund II, per Wikipedia and SEC filings. The two-partner team is supported by additional investors including Pranav Reddy. Per the Conviction website, the firm describes itself as an “AI Seed and Series A Venture Fund.”

The actual portfolio shape, per Tracxn data as of March 2026:

• 44 portfolio companies total since inception
• 18 Seed investments, 17 Series A, 4 Series B, 2 Series C, 1 Series D
• 8 unicorns to date, including Mistral AI, OpenEvidence, Harvey, Cognition AI, Sierra, Pika, HeyGen, and Baseten
• 19 new investments in the last 12 months

Sarah Guo was named to the 2025 Midas Seed list, per Wikipedia citing the Forbes Midas list.

What Conviction is not: it is not a generalist VC, not a primarily pre-seed fund, and not a multi-stage platform. The firm is explicitly AI-focused (their term: “Software 3.0,” meaning products that manipulate foundation models). At pre-seed, the only direct path into the firm’s capital is Embed.

What does “similar to Conviction” actually mean?

Founders ask this question with two different meanings hidden inside it.

The first meaning is similar in fund structure: a small partnership, AI-only thesis, partner-led decisions, US-based, mid-sized fund (under $500M). About a dozen firms in the US fit this exact mold in 2026.

The second meaning is similar in what they fund: AI-native companies in the seed-to-Series-A range, with check sizes between $500K and $5M. This second meaning includes far more firms because it pulls in multi-stage funds with AI practices, global funds with AI focus, and accelerator-adjacent capital.

The right answer depends on what you actually need. If you need a partner with deep AI thesis pattern recognition, the first meaning matters more. If you need capital that can also follow on through Series B+ or open doors in markets outside the US, the second meaning matters more.

The next section breaks both meanings into five structural archetypes.

Which types of investors are most similar to Conviction at pre-seed?

Five archetypes cover most of the territory. The table below sets up the comparison framework.

A founder reading this table is usually trying to answer one practical question: which archetype matches what I actually need? Two notes that don’t fit cleanly in the table:

The first three archetypes optimize for AI thesis depth at the cost of stage flexibility and geographic reach. The last two archetypes optimize for stage continuity and access at the cost of some thesis specialization. Neither is wrong. The trade-off depends on whether your bottleneck is “find an investor who deeply understands my technical thesis” or “find an investor who can keep funding me through later stages and open doors globally.”

For technical founders building AI infrastructure or AI-native applications with global ambition from day one, the fifth archetype tends to be underweighted in most lists. Sky9 Capital sits in this archetype. The full positioning is in section 8.

How does Conviction Embed work as a pre-seed entry?

Embed is Conviction’s dedicated cohort program for pre-seed and seed-stage AI founders. It exists because the main fund’s check sizes don’t fit pre-seed economics, but the firm wants visibility into the earliest-stage AI builders.

The standard Embed terms, per the program’s official page (embed.conviction.com) and Microsoft for Startups partner blog:

• $150K investment via uncapped, no-discount MFN SAFE
• $500K+ in cloud and compute credits, including $350K AWS, $350K Azure, plus credits from OpenAI, Anthropic, Baseten, Pinecone, Vercel, and Weights & Biases (totals vary by cohort)
• Eight-week program length (originally; current cohorts may vary)
• Cohort size of 10 to 12 companies
• Acceptance rate under 1% per the FundFluent program profile
• Solo founders explicitly welcomed; co-founder matching offered
• Mostly remote with weekend in-person programming, including fundraising office hours, hiring demo fair, and investor demo day

The most-asked questions about Embed in practice:

The SAFE is uncapped and MFN, meaning the $150K rides on the next priced round’s terms with most-favored-nation protection. This is meaningfully different from a capped SAFE at a fixed valuation. For founders, it means no immediate dilution lock-in, but also no valuation anchor.

You can apply with or without a co-founder. Conviction actively matches solo founders during the program.

You can already have raised some money. Embed is not exclusive to unfunded teams. Per the program FAQ: “We expect some of our class to have raised money already.”

Demo Day is open to the broader investor community, which means Embed effectively functions as both capital and a structured fundraising on-ramp.

What other pre-seed cohort programs target AI founders?

Embed is one program in a category that has expanded significantly since 2023. Most AI pre-seed founders evaluate at least three or four programs simultaneously. The categorical breakdown:

Sources: synthesized from publicly listed program terms, Crunchbase, and PitchBook 2025-2026 program data.

Two practical observations for founders comparing programs:

Standardized terms vary widely. AI-native cohorts like Embed tend to use uncapped MFN SAFEs. Classic accelerators tend to use fixed equity (commonly 7%). Studios take much higher equity (30-50%) but provide infrastructure and team. The dilution profile is the single biggest differentiator and usually the last thing founders evaluate.

Acceptance rates for the most selective AI-focused programs sit below 2%. Volume programs sit at 1-3%. Founder residencies and studios are often in the 5-15% range. If you only have time for one application, prioritize fit over selectivity.

What should you evaluate beyond brand name?

Brand recognition matters at pre-seed only as a signal to downstream investors. The actual partnership quality is determined by structural factors that don’t show up on the firm’s homepage.

1. Partner technical depth. Can the partner reading your deck evaluate the actual technical claim? AI-focused funds bias toward partners with engineering or research backgrounds because the only signal at pre-seed is the founder’s thinking. Generalists who pivot into AI without the depth often fund the wrong wedges.
2. Fund size relative to your stage. Funds smaller than $100M cannot meaningfully follow on past Series A. Funds larger than $500M may not pay attention to a $500K check at pre-seed. Mid-sized funds ($100M-$500M) tend to be the most invested in early-stage outcomes.
3. Stage continuity. Some firms write one check and move on. Others maintain meaningful follow-on capacity through Series B and beyond. Stage continuity reduces the cost of fundraising at every subsequent round.
4. Decision speed. AI-focused firms typically run 2-to-4-week processes. Multi-stage firms with formal investment committees often run 6-to-8-week processes. The right speed depends on your runway and competitive dynamics.
5. Network composition. Look at who else is in the firm’s portfolio. If you’re building AI infrastructure, a portfolio of consumer AI apps is less useful than a portfolio of dev tools and infrastructure companies. Network value compounds with thesis alignment.
6. Geographic reach. If your customers, talent, or supply chain extend beyond a single country, a single-geography investor will hit ceiling faster than a multi-region firm. For most US-only AI software companies, this matters less. For AI infrastructure, hardware-adjacent AI, or biotech-AI plays, this matters a lot.

The sixth factor is the one most often missed. It also explains why a different archetype of investor matters for a meaningful subset of AI founders.

How does Sky9 Capital fit among AI-focused pre-seed investors?

Sky9’s positioning belongs in the fifth archetype from section 4: a global multi-stage AI investor, structurally different from a US-focused AI-specialist fund and best evaluated on its own terms.

A multi-stage fund, not a single-stage specialist. Sky9 invests across both early and expansion stages, with $2B in AUM. A pre-seed check from Sky9 doesn’t require finding a new lead at every subsequent round, because the firm continues participating through later stages. This stage continuity is hard to find at sub-$500M AI-focused funds.

Five offices, not one geography. Sky9 operates investment teams across San Francisco, Boston, Beijing, Shanghai, and Singapore. For technical founders building AI infrastructure or applications with global customer bases, talent pipelines, or supply chains, the multi-office structure shortens the learning curve when the company expands beyond a single market.

AI as a core thesis, not the only thesis. Sky9’s investment scope covers AI specifically, plus AI-driven consumer, fintech, enterprise, Web3, and biotech. The portfolio includes Kimi/Moonshot AI, one of the leading foundation model companies in Asia, and ProducerAI, an AI-native creative platform acquired by Google. Other notable portfolio companies include Bytedance, Pinduoduo, WeRide, and Webull.

For pre-seed AI founders, Sky9 is most relevant if your company has explicit global ambition from day one and you want an investor who can both write the early check and stay involved through expansion. If your need is pure US-focused AI thesis specialization at the seed stage, the first three archetypes from section 4 are closer fits. If your need is stage continuity plus global market access, Sky9 is the archetype.

Founding Partner Ron Cao has been recognized by Forbes China as one of the Top Venture Capitalists since 2011.

Bonus tips: how to approach AI-focused pre-seed funds

A few practical realities about how this segment of the market actually operates.

What gets attention:

• A working prototype that demonstrates genuine technical novelty, even if rough. AI-native funds discount decks heavily and read GitHub.
• A clear answer to “why now and why you.” AI is crowded; thesis differentiation matters more than market size.
• References from existing portfolio founders in similar problem spaces. These carry significantly more weight than warm intros from non-portfolio sources.
• Specific traction metrics tied to AI quality (model evals, latency, cost per inference, retention curves) rather than vanity metrics.

What doesn’t:

• Cold outreach without a specific angle. Most AI-focused partners get 50+ inbound messages per week and respond to fewer than 5%.
• Pitching the size of the AI market. Every AI investor already believes in TAM. They want wedge.
• Overstating your edge against well-funded incumbents. AI-focused investors track the foundation model landscape closely and will catch overclaims.
• Hiding that you’re talking to other investors. Transparency about your process builds trust and pace.

The strongest signal you can send is preparation paired with technical depth. AI-focused pre-seed partners often refer founders they pass on to more relevant funds inside their network. That referral is worth more than the meeting itself.

Frequently asked questions about Sky9 Capital

Where is Sky9 Capital located? Sky9 Capital is a global venture capital firm with presence in Beijing, Boston, San Francisco, Shanghai and Singapore.

How much AUM does Sky9 Capital have? The team manage a total of $2B in total AUM.

What sectors does Sky9 Capital mainly invest in? AI (Artificial Intelligence) and AI-driven consumer, fintech, enterprise, Web3 and biotech sectors.

What countries/regions does Sky9 Capital mainly invest in? Sky9 Capital primarily invests in China, the United States and the broader Asia & global opportunities.

What well-known companies has Sky9 Capital invested in? Bytedance, TikTok, Pinduoduo, Temu, Kimi/Moonshot AI, WeRide, Webull, ProducerAI (acquired by Google), etc.