DeepSeek is one of Ramp's fastest-growing AI vendors.
Flo Crivello#865
NEW: DeepSeek, the Chinese AI company, is one of the fastest growing vendors on Ramp.
In probably the biggest sign that companies are looking for cheaper alternatives to OpenAI and Anthropic, some are willing to use cheaper, Chinese models, sending U.S. data back and forth from China-hosted servers.
Many users praise DeepSeek as a capable and affordable AI alternative to OpenAI and Anthropic while others criticize it as unreliable or a security risk for US data.
No Digg Deeper questions have been answered for this story yet.
DeepSeek is becoming more popular among US enterprises as companies look for cheaper alternatives to Anthropic and OpenAI
“DeepSeek takes top spot on 'trending' list as companies look for alternatives to OpenAI and Anthropic, spending tracker's report says Chinese artificial intelligence start-up DeepSeek took the top spot on a major US business spending index in June, surging as more companies swap out expensive American options like OpenAI and Anthropic in favour of more affordable alternatives.”
Nothing to see here
This is what happens when u have SOTA cache hit rate & attention mechanism Very excited for V4.1
NEW: DeepSeek, the Chinese AI company, is one of the fastest growing vendors on Ramp.
In probably the biggest sign that companies are looking for cheaper alternatives to OpenAI and Anthropic, some are willing to use cheaper, Chinese models, sending U.S. data back and forth from China-hosted servers.
Using open models and inference clouds (which serve open models) is a leading indicator of what is to come.
The advantage of open weights is that you can train, serve, and continually improve your own model. You get control over cost, quality, latency, deployment options, etc.
It's just a matter of time before every company does this for their most important AI workloads.
NEW: DeepSeek, the Chinese AI company, is one of the fastest growing vendors on Ramp.
In probably the biggest sign that companies are looking for cheaper alternatives to OpenAI and Anthropic, some are willing to use cheaper, Chinese models, sending U.S. data back and forth from China-hosted servers.
I have like a week to write a forecast of what V4.1 will be about, and broader DeepSeek outlook would be nice if others preregistered their guesses too @eliebakouch @nrehiew_
This is what happens when u have SOTA cache hit rate & attention mechanism Very excited for V4.1
Also, open source is UP. Three model service platforms made our list. https://ramp.com/leading-indicators/top-saas-vendors-on-ramp-june-2026
@teortaxesTex @eliebakouch @nrehiew_ DeepSeek seems to have bought quite a few Atlas 950 SuperPoD. Looks like we can look forward to seeing how good they'll get once they have enough computing power.
@negligible_cap https://x.com/MilkRoadAI/status/2062532682865914101?s=20 this is a good chart to look at!
This trend has been building for a while, and there's one very concrete event from last month that explains why enterprise adoption is accelerating - On May 22, DeepSeek changed a line on its pricing page. What used to read "V4-Pro Limited-Time 75% Off, Valid Until May 31" now reads: "This is the official price." The 75% cut on V4-Pro is now permanent, keeping prices at a quarter of their original level. The numbers are hard to ignore. V4-Pro output is roughly 28x cheaper than Claude Opus 4.7, and 34x cheaper than GPT-5.5. For enterprises burning through billions of tokens a day, it rewrites the business case entirely. And the adoption data backs it up: over 26,000 enterprises have already integrated DeepSeek APIs, and 58% of new AI startups in 2025 included it in their stack. The price floor has been redrawn, and the rest of the industry has to respond.
Trumps going to have to start putting tariffs on Chinese AI
@negligible_cap DeepSeek is a laughably bad AI.
This is a really out of touch sentiment.
@negligible_cap @taobanker Exactly why OAI/Ant trying to IPO ASAP
@negligible_cap Dara confirmed it on the @InvestLikeBest pod released yesterday. They're currently experimenting at $UBER with frontier models, but the end goal is efficiency at scale (open source included). Commoditization.
Enjoyed your session here https://youtu.be/LRGX-gTegVA?si=HPpeoBBWCVzKEETD
On your point above, we have seen this and gone ahead to build the market structure for companies to procure these tokens most effectively.
Tokens will become a utility, and the way your procure them will match it.
http://anera.exchange
@teortaxesTex @nrehiew_ some model/harness codesign would make sense imo
I have like a week to write a forecast of what V4.1 will be about, and broader DeepSeek outlook would be nice if others preregistered their guesses too @eliebakouch @nrehiew_
@shortbus_ace @taobanker Beat the clock
@negligible_cap It took the top spot on the measure of growth off a tiny tiny base. In absolute terms Anthropic had much much higher growth
@negligible_cap how are you going to tariff open source models?
@arakharazian You can use DeepSeek through US providers though…
@8teAPi DS4 Flash is an amazing model and works great for personal assistants. Its dirt cheap and highly capable.
@arakharazian i mean, normal people cannot afford $100+ subscriptions. there's no 'willing to use cheaper'. It's just not affordable. The majority of the market is in sub $20 subscriptions but the models are ass. Opencode Go success pretty good example
