Jochem πŸ§‘β€πŸš€ (@Jochem_In_Space) header
Jochem πŸ§‘β€πŸš€ (@Jochem_In_Space) profile picture

AI Ops @AxeCompute & Not your average free diver 🀿 πŸ“ Views are my own

Finding my way 🧭 in space 🌌
Joined April 2015
Photo by Jochem πŸ§‘β€πŸš€ (@Jochem_In_Space)Photo by Jochem πŸ§‘β€πŸš€ (@Jochem_In_Space)Photo by Jochem πŸ§‘β€πŸš€ (@Jochem_In_Space)Photo by Jochem πŸ§‘β€πŸš€ (@Jochem_In_Space)Photo by Jochem πŸ§‘β€πŸš€ (@Jochem_In_Space)
Jochem πŸ§‘β€πŸš€ retweeted
Axe Compute (NASDAQ: $AGPU) will host its Q3 2026 Financial Results Conference Call on November 17 at 8:30 AM ET. Join our leadership team for financial results and business updates. Dial-in: +1 301 715 8592 ID: 88188227988 Passcode: 457947
2
3
2
39
54,351
We have a new home base! Miami here we come!
Axe Compute (NASDAQ: $AGPU) is making Miami its new global headquarters. The move strengthens our presence in a major international business and financial hub as we expand our global AI infrastructure operations, partnerships and enterprise relationships. #AxeCompute #AIInfrastructure
30
Jochem πŸ§‘β€πŸš€ retweeted
Last week Axe Compute rang the @NasdaqExchange Opening Bell in Times Square. A milestone for the team, and a marker on the way to something bigger. The building is just beginning, and we’re here to define the next generation of AI infrastructure. We’re just getting started. $AGPU #AIInfrastructure
13
26
2
356
324,885
This is another article I am inspired by as it shows how AI is used to simulate even the toughest B2B environments.
A factory digital twin runs three GPU workloads at once: rendering, physics, and AI. All three compete for the same GPU memory. NVIDIA’s engineering-twin blueprint needs 16 GB to show pre-computed results, and two 40 GB GPUs to run the AI physics live. BMW Group already runs twins of 30+ production sites, cutting collision checks from almost four weeks to three days. Our new article explains what a twin is for AI builders, and how to size the cluster for the live twin, not the demo. Full article below.
23
Jochem πŸ§‘β€πŸš€ retweeted
teapaw.org/i/broadcasts/1kJzDPbvd… We’re live from Times Square. Axe Compute (NASDAQ: $AGPU ) is ringing the Nasdaq Opening Bell now. @NasdaqExchange
3
2
16
1,098
Jochem πŸ§‘β€πŸš€ retweeted
Video games are becoming a serious training ground for robots. General Intuition raised a $133.7 million seed round on that idea. Gameplay records an action and exactly what followed, and games already encode what makes physical space hard to model: objects persist, occlude each other, and move under causal rules. It points at a bigger shift. Systems that act in the world need a persistent model of it, and that turns bursty inference into sustained, stateful GPU workloads. Axe Compute lets teams pick the platform, region, and term, on single-tenant bare metal with zero egress fees. That matters when a pipeline moves petabytes of video between stages. Read our full piece on spatial AI. Link in the comments.
1
4
13
637
Jochem πŸ§‘β€πŸš€ retweeted
What happens when data center growth slows down? When data center capacity can’t keep pace with demand, access to the compute that powers AI can become a bottleneck for research, discovery and innovation. Chris Miglino's answer on @FoxBusiness was direct: "When we slow down AI infrastructure, we slow down the cure to cancer." AI infrastructure is becoming critical to research, innovation and building what comes next. $AGPU
4
7
2
24
12,599
Jochem πŸ§‘β€πŸš€ retweeted
Axe Compute has received more than $184M in customer prepayments and signed over $266M in new AI infrastructure contracts since the Q2 earnings call. Total contracted value signed over the past seven months now exceeds $3.45B. Full release: investors.axecompute.com/new… $AGPU
7
6
6
37
23,040
Jochem πŸ§‘β€πŸš€ retweeted
A GPU's peak FLOPS number describes ideal lab conditions, and it is a fine way to compare chip generations. It does not tell you what your model will get once a rack is live. Google's PaLM paper reports 46.2% Model FLOPs utilization at 540B parameters, against 21.3% for GPT-3, from two of the most capable labs in AI. Google's TPU v4 research measured about 60% of peak on a real training run. If labs at that scale see swings like that, expect at least as much variance on any rack built for a single workload. This is why companies want to design their own clusters. Axe Compute Build designs clusters to your spec: bare metal, single tenant, and backed by enterprise-grade SLAs. Full analysis: axecompute.com/how-to-benchm…
1
4
1
10
3,118
Jochem πŸ§‘β€πŸš€ retweeted
Inside the build in Columbus, Georgia. Axe Compute President Kyle Okamoto gives a look at the work underway to bring our 288-node @nvidia Blackwell B300 cluster online. Containment is being finished. Cabling is underway. And the cluster is designed to scale to roughly 3x its current size in the coming months. #AIInfrastructure #AIDatacenter #NVIDIA #Blackwell
19
20
82
4,842
Jochem πŸ§‘β€πŸš€ retweeted
AI compute is getting tokenized, and @binance just listed these names, including @AxeCompute as a rising neocloud. This listing shows the market's continued recognition of Axe Compute's GPU-as-a-Service business in 2026. Congrats to Axe Compute on the AGPU listing on Binance Futures. binance.com/en-ZA/support/an…
10
20
2
516
24,373
Jochem πŸ§‘β€πŸš€ retweeted
Axe Compute is getting noticed. @TheStreet takes a closer look at our AI infrastructure transformation, customer contracts, GPU deployments, and the growing demand for dedicated AI compute. Featuring insights from CEO Christopher Miglino and President Kyle Okamoto. Read the full story: thestreet.com/investing/stoc… #AxeCompute #AIInfrastructure #AI
6
3
19
1,801
New ad by the Norwegian Consumer Council about the declining lifespan of products.
407
10,807
1,066
47,169
2,579,149
$AGPU is ready to dedicate everything to that one big goal!
Today we announced the sale of Helomics, our Pittsburgh-based AI oncology business, to DataMEDS AI (NASDAQ: MEDS). With this step complete, our team, capital, and operations are dedicated entirely to one goal: meeting the accelerating demand for AI compute. We are fully committed to delivering best-in-class AI compute infrastructure, giving enterprises choice across hardware, geography, and deployment speed through Axe Compute Access and Axe Compute Build. Read the full announcement: investors.axecompute.com/new…
1
3
177
Why voice AI teams will eventually require their own dedicated clusters, and @AxeCompute can built to spec for them! $AGPU
According to research on voice AI latency, callers expect a response within roughly 300 milliseconds, the natural pause length in ordinary human conversation. When that line is crossed the conversation feels broken, even when the answer is correct. Other research on voice AI shows that the median time-to-first-token is near 500 milliseconds, which does not include speech recognition and synthesis. Voice AI teams are looking for their own dedicated infrastructure to fill that gap. Axe Compute Build stands up dedicated clusters to the exact spec a voice workload needs, bare metal and single tenant. Full analysis: axecompute.com/voice-ai-real… Reserve capacity at portal.axecompute.com
1
190
Jochem πŸ§‘β€πŸš€ retweeted
Training is an event. Inference is a utility. That distinction defines infrastructure planning. Gartner reports global inference spending hit $23.3 billion in 2025, overtaking the $19 billion spent on training, according to CIO Dive. Deloitte's November estimate goes further: inference already accounts for half of all AI compute in 2025, rising to two thirds in 2026. At CES 2026, Lenovo's Ashley Gorakhpurwalla told Computerworld his company now forecasts an 80/20 split favoring inference, a full reversal from the last buildout cycle. Training has a start date and an end date. Inference grows with every new customer. Enterprises that size infrastructure with a training mindset overbuild for peak and underbuild for sustained load. Axe Access provisions the latest GPUs in as fast as 48 hours across numerous global locations, built for exactly this kind of continuous, usage-driven demand. Axe Build stands up dedicated capacity, bare metal and single tenant, when that continuous load needs its own dedicated footprint. Full analysis: axecompute.com/inference-shi… Reserve capacity at portal.axecompute.com.
2
2
9
697
Jochem πŸ§‘β€πŸš€ retweeted
Thank you, @weka. We’re proud to bring together technology from trusted providers like WEKA with @nvidia Blackwell B300 compute as we build production-ready AI infrastructure projects designed for demanding, large-scale workloads. #AGPU #AIFactories #AIinfrastructure #AxeBuild
🀝 Proud to partner with @AxeCompute on a large-scale AI factory in Georgia, featuring 288 @nvidia Blackwell Ultra GPU nodes and 20+ PB of high-speed WEKA storage. πŸŽ₯ See it in action: weka.ly/469uabB
4
5
20
1,263
Jochem πŸ§‘β€πŸš€ retweeted
Eight GPUs. One system. One massive computing environment. That’s what makes the infrastructure behind modern AI so powerful. Axe Compute President Kyle Okamoto goes inside our @nvidia Blackwell B300 cluster and explains how InfiniBand networking allows GPUs across the facility to work together for high-speed, high-scale AI workloads.
4
3
18
1,471
Jochem πŸ§‘β€πŸš€ retweeted
$317M in customer prepayments. ~$3B in contracted infrastructure. Axe Compute $AGPU CEO Christopher Miglino tells @Investingcom how the company is matching capital to signed customer commitments as it scales its AI infrastructure buildout. β€œContracts first, capital against contracts, deployment against capital.” Read the full exclusive: investing.com/news/stock-mar… #AIInfrastructure #AI
7
8
40
1,321