Resolvemos por meio deste blog estabelecer um canal direto de comunicação do grupo executivo da ODATA com nossos clientes, prospects e parceiros. É também uma forma de conhecer um pouco melhor nossa visão, forma de pensar e um pouco da cultura ODATA.
My good friend Mauricio Ruiz published a post linking to the following article “ Why do Latin Americans pay more for cloud computing?“ and it got me thinking. The article raises some important points, but I think it misses other causes that are no less important.
One of the main reasons for the cost of cloud to be higher in Brazil vs the US is obvious, and we can draw the parallel to an iPhone. Brazil has the most expensive iPhone in the world, mostly due to very high import, sales and all sorts of taxes that fall on a fully imported (not assembled here) electronic device. Those exact same taxes apply to servers. Large cloud providers such as Amazon, Microsoft, IBM Softlayer and others build their own servers, working closely with OEMs and chip providers that create specific products for their unconventional workloads; they consider it to be their “secret sauce”. Therefore, their cloud structures (servers, switches, storage, etc.) come, literally, in containers, ready assembled in racks that are rolled into local Data Centers. Therefore, they will pay full import taxation on the equipment. The author of the article alludes to an opportunity for startups to assemble locally and bridge the gap. I disagree, it´s not so simple. Even if you assemble locally, you need volume, significant volume, and even if you have a strong market demand, this volume will be peanuts compared to that of AWS, Azure et al. Let´s assume that one does manage to assemble locally, none of the parts are made here. Almost every single component needs to be shipped from Taiwan. The most expensive part of the servers is silicon (CPU, memory, chipset) which is 100% imported and thus incurs in the same import taxes. Tax advantages for assembling locally will always be trumpeted by high volume efficiencies.
Besides these costs, there are other not so obvious culprits, that are crucial and get factored into the cloud provider´s price to the final customer. Locally, we refer to it as “Brazil Cost” or “Custo Brasil”. It includes all sorts of costs incurred by companies to operate in a country that the basic infrastructure is scarce and inefficient. For example, fiber and electricity for Data Centers. In US, DCs are built in remote areas where land is cheap and there are advantages to developers, such as low cost of energy, tax breaks, cool climate, etc. In Brazil and in most Latin American countries, that is not possible. To put a Data Center in a remote area is a losing proposition, since large capacity fiber is only available on the densest urban centers or along major roads where the real estate is most expensive. Consequently, to connect this remote DC to the world, either the DC provider, or the fiber company will need to incur in significant investments. Multiply that by 10 or 15 carriers that each of these sites have and you arrive at a large bill.
Power utility is another deterrent, since existing infrastructure lacks serious upgrades and the investment to connect to high tension lines (and consequently become a major customer of these utilities) the client needs to make all the investment including building its own substation and distribution lines to the power company. This becomes a major cost to Data Center providers, one that in most countries does not exist. I was surprised, while in a visit to an European country, inside a very large DC with 17Mw of power, I asked if they had to build anything to connect to the power utility. They laughed, saying that the incumbent there was so happy to have such a big client that they gave them 2 separate connections coming from different substations in their grid.
And these are the direct, more obvious ones. When we look at the value chain, inefficiencies in the price of energy itself (Brazil just came out of a major drought, sending electricity prices sky-high), ports, transportation, security, construction, etc. all trickle down.
But the very good news is that even with all these issues, cloud is just starting to penetrate in the corporate budgets in Latin America, the growth curves are exponential, and with high volume, there is an inexorable tendency that a large part of these gains are passed directly to the final customer. Likewise, on the Data Center side, we are working tirelessly to service the cloud providers with the most efficient and competitive offer possible, since scale also brings substantial efficiencies that are passed to our direct customers. In time, I believe that the price gap of cloud services in Latin America versus that of more developed countries will reduce significantly.