Technology
Charting the Depths: The World of Subsea Cables
Charting the Depths: The World of Subsea Cables
Data may be stored in the “cloud,” but when it comes to sending and receiving data, a lot of that action is actually happening along the depths of the ocean floor.
Hidden beneath the waves, these subsea cables account for approximately 95% of international data transmission.
These maps, by Adam Symington, use information from TeleGeography to show the distribution of subsea cables around the planet.
Wired for Connectivity
It’s estimated that there are nearly 1.4 million kilometers (0.9 million miles) of submarine cables in service globally. They ensure emails, content, and calls find their way, linking colossal data centers and facilitating worldwide communication.
Currently, there are 552 active and planned submarine cables:
Submarine cables harness fiber-optic technology, transmitting information via rapid light pulses through glass fibers. These fibers, thinner than human hair, are protected by plastic or even steel wire layers.
Cables usually have the diameter of a garden hose, but often with added armor near the shore. Coastal cables are buried under the seabed, hidden from view on the beach, while deep-sea ones rest on the ocean floor.
Length varies widely, from the 131-kilometer CeltixConnect cable, connecting Dublin, Ireland, and Holyhead, UK, to the sprawling 20,000-kilometer Asia America Gateway cable, connecting San Luis Obispo, California, to Hawaii and Southeast Asia:

Asia America Gateway. Image: TeleGeography
With the current technology, cables are designed to last 25 years at least but are often replaced because of damage. Nearly two-thirds of cable damage is caused by fishing vessels and ships dragging anchors.
The Bottom Line
Traditionally dominated by telecom carriers, the makeup of the subsea cable market has shifted over more recent decades. Tech giants like Google, Facebook, Microsoft, and Amazon now heavily invest in new cables.
With data demand surging, at least $10 billion is expected to be invested in subsea cables worldwide between 2022 and 2024, driven by cloud service providers and content streaming platforms.
Even with the growth of satellites in telecom, cables still can carry far more data at a much lower cost than satellites. In fact, according to TeleGeography, satellites account for less than 1% of all U.S. international capacity.

This article was published as a part of Visual Capitalist's Creator Program, which features data-driven visuals from some of our favorite Creators around the world.
Technology
Visualizing AI vs. Human Performance in Technical Tasks
AI systems have seen rapid advancements, surpassing human performance in technical tasks such as advanced math and visual reasoning.

AI vs. Human Performance in Technical Tasks
This was originally posted on our Voronoi app. Download the app for free on iOS or Android and discover incredible data-driven charts from a variety of trusted sources.
The gap between human and machine reasoning is narrowing—and fast.
Over the past year, AI systems have continued to see rapid advancements, surpassing human performance in technical tasks where they previously fell short, such as advanced math and visual reasoning.
This graphic visualizes AI systems’ performance relative to human baselines for eight AI benchmarks measuring tasks including:
- Image classification
- Visual reasoning
- Medium-level reading comprehension
- English language understanding
- Multitask language understanding
- Competition-level mathematics
- PhD-level science questions
- Multimodal understanding and reasoning
This visualization is part of Visual Capitalist’s AI Week, sponsored by Terzo. Data comes from the Stanford University 2025 AI Index Report.
An AI benchmark is a standardized test used to evaluate the performance and capabilities of AI systems on specific tasks.
AI Models Are Surpassing Humans in Technical Tasks
Below, we show how AI models have performed relative to the human baseline in various technical tasks in recent years.
Year | Perfomance relative to the human baseline (100%) | Task |
---|---|---|
2012 | 89.15% | Image classification |
2013 | 91.42% | Image classification |
2014 | 96.94% | Image classification |
2015 | 99.47% | Image classification |
2016 | 100.74% | Image classification |
2016 | 80.09% | Visual reasoning |
2017 | 101.37% | Image classification |
2017 | 82.35% | Medium-level reading comprehension |
2017 | 86.49% | Visual reasoning |
2018 | 102.85% | Image classification |
2018 | 96.23% | Medium-level reading comprehension |
2018 | 86.70% | Visual reasoning |
2019 | 103.75% | Image classification |
2019 | 36.08% | Multitask language understanding |
2019 | 103.27% | Medium-level reading comprehension |
2019 | 94.21% | English language understanding |
2019 | 90.67% | Visual reasoning |
2020 | 104.11% | Image classification |
2020 | 60.02% | Multitask language understanding |
2020 | 103.92% | Medium-level reading comprehension |
2020 | 99.44% | English language understanding |
2020 | 91.38% | Visual reasoning |
2021 | 104.34% | Image classification |
2021 | 7.67% | Competition-level mathematics |
2021 | 66.82% | Multitask language understanding |
2021 | 104.15% | Medium-level reading comprehension |
2021 | 101.56% | English language understanding |
2021 | 102.48% | Visual reasoning |
2022 | 103.98% | Image classification |
2022 | 57.56% | Competition-level mathematics |
2022 | 83.74% | Multitask language understanding |
2022 | 101.67% | English language understanding |
2022 | 104.36% | Visual reasoning |
2023 | 47.78% | PhD-level science questions |
2023 | 93.67% | Competition-level mathematics |
2023 | 96.21% | Multitask language understanding |
2023 | 71.91% | Multimodal understanding and reasoning |
2024 | 108.00% | PhD-level science questions |
2024 | 108.78% | Competition-level mathematics |
2024 | 102.78% | Multitask language understanding |
2024 | 94.67% | Multimodal understanding and reasoning |
2024 | 101.78% | English language understanding |
From ChatGPT to Gemini, many of the world’s leading AI models are surpassing the human baseline in a range of technical tasks.
The only task where AI systems still haven’t caught up to humans is multimodal understanding and reasoning, which involves processing and reasoning across multiple formats and disciplines, such as images, charts, and diagrams.
However, the gap is closing quickly.
In 2024, OpenAI’s o1 model scored 78.2% on MMMU, a benchmark that evaluates models on multi-discipline tasks demanding college-level subject knowledge.
This was just 4.4 percentage points below the human benchmark of 82.6%. The o1 model also has one of the lowest hallucination rates out of all AI models.
This was major jump from the end of 2023, where Google Gemini scored just 59.4%, highlighting the rapid improvement of AI performance in these technical tasks.
To dive into all the AI Week content, visit our AI content hub, brought to you by Terzo.
Learn More on the Voronoi App 
To learn more about the global AI industry, check out this graphic that visualizes which countries are winning the AI patent race.
-
Misc2 weeks ago
Mapped: The Most Popular Beer in Each U.S. State
-
Maps4 weeks ago
Mapped: How Much a 24-Pack of Beer Costs in Each U.S. State
-
Maps3 weeks ago
Mapped: Which U.S. States Import the Most from China?
-
Money3 weeks ago
Ranked: Daily Incomes of the Richest and Poorest in 25 Countries
-
Automotive4 weeks ago
Ranked: The Cheapest Car Brands to Own and Maintain in the U.S.
-
Money2 weeks ago
Mapped: The Most Taxed States in America
-
United States3 weeks ago
Mapped: The Best Selling Vehicle in Every U.S. State in 2024
-
Business1 week ago
Timeline: Visualizing America’s Oldest Companies